Studies on the mechanism and physiological role (s) of the interaction of ataxin-3 with tubulin

Ataxin-3 (AT3) is a protein consisting of an N-terminal globular Josephin domain and an unstructured C-terminal region that contains a stretch of consecutive glutamines. When its length is beyond a critical threshold, it triggers an inherited neurodegenerative disease, spinocerebellar ataxia type 3. The pathology results from protein misfolding and intracellular accumulation of fibrillar amyloid-like aggregates. Plenty of work has been carried out to elucidate the protein’s physiological role(s), which has shown that AT3 is multifunctional protein. It acts as transcriptional repressor and also carries out its normal function on protein surveillance pathways. In fact, AT3 interacts with RAD23, which shuttles ubiquitinated protein to proteasome, and also with VCP that seems to be involved in the degradation pathway ubiquitin-proteasome dependent and in the retrotranslocation of ERAD substrates. In addition, a recent report suggests that it participates in sorting misfolded protein to aggresomes. Furthermore, it was shown that AT3 binds ubiquitin chains through its ubiquitin-binding motifs and also has, in the Josephin domain, the catalytic triad of thiol proteases that in this protein sustains ubiquitin cleavage. Since a thorough understanding of the protein’s physiological role(s) requires the identification of all the molecular partners interacting with AT3, we pursued this goal by taking advantage of two-dimensional chromatography coupled to tandem mass spectrometry. We found that different AT3 constructs, including the sole Josephin domain, bound α- and β- tubulin from soluble rat brain extracts. Coimmunoprecipitation experiments confirmed this interaction. Also, normal AT3 overexpressed in COS-7 cultured cells partially colocalized with microtubules, whereas an expanded variant only occasionally did so, probably due to aggregation. Furthermore, by surface plasmon resonance (SPR) we determined a dissociation constant of 50–70 nM between AT3 and tubulin dimer, which strongly supports the hypothesis of a direct interaction of this protein with microtubules in vivo. Since misfolded protein is sorted to aggresome via microtubules under conditions where proteasome is overloaded or its function compromised, our finding suggests an involvement of AT3 in directing aggregated protein to aggresome. In further experiments, we assessed by SPR the binding affinity of AT3 with either Lys48-linked or Lys63-linked polyubiquitin chains. We observed that either chains were bound with high affinity by AT3. However, we also found that binding of Lys48-polyubiquitin and tubulin were mutually exclusive, whereas Lys63-polyubiquitin and tubulin could be bound simultaneously. Protein Lys48- and Lys63-polyubiquitination are signals that direct to proteasome and aggresome, respectively. Thus, our findings point to a model whereby AT3 sorts tagged protein to either goals, depending on the type of ubiquitin bound. In the former case, AT3 cannot bind to tubulin so it is prevented from being directed to aggresome; in the latter it sorts misfolded protein to aggresome. In keeping with this hypothesis, we also found that AT3 binds to histone deacetylase 6 (HDAC6) with nanomolar affinity. Actually, the latter protein is indispensable for aggresome formation, as it also binds Lys63-polyubiquitylated proteins and interacts with the dynein complex, which transports misfolded protein cargo to aggresomes via microtubules. Finally, we also investigated the mode of interaction between AT3 and tubulin by producing several truncated AT3 variants, and determining their affinity towards tubulin. These studies highlight the occurrence of three different tubulin-binding sites: in the Josephin domain, in the disordered stretch upstream of the polyQ tract, and at the C-terminus, at the very end of the polypeptide chain.